GoogleModel generates empty text parts that Vertex AI rejects with 400 error #2032
Description
Initial Checks
- I confirm that I'm using the latest version of Pydantic AI
- I confirm that I searched for my issue in https://github.com/pydantic/pydantic-ai/issues before opening this issue
Description
When using GoogleModel with Vertex AI, requests fail with a 400 error because pydantic_ai generates empty text parts {'text': ''} that Google's API rejects.
Error Message
400 Bad Request. {
"error": {
"code": 400,
"message": "Unable to submit request because it must have a text parameter. Add a text parameter and try again. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini",
"status": "INVALID_ARGUMENT"
}
}Root Cause
In pydantic_ai/models/google.py, the _map_messages() method at lines 362-364 adds empty text parts as a defensive measure:
# Google GenAI requires at least one part in the message.
if not message_parts:
message_parts = [{'text': ''}]However, Google's Vertex AI API strictly validates the text parameter and rejects empty strings, requiring non-empty text content.
Reproduction
This occurs when:
- Messages have no content after processing
- Message parts arrays become empty during conversion
- Certain edge cases in message handling
Impact
- Affects all users of
GoogleModelwith Vertex AI - Causes complete request failures with 400 errors
- Blocks streaming and non-streaming requests
Suggested Fix
Replace the empty string with a minimal valid text:
# Instead of:
if not message_parts:
message_parts = [{'text': ''}]
# Use:
if not message_parts:
message_parts = [{'text': ' '}] # Single spaceOr add provider-specific handling for Google's stricter validation requirements.
Temporary Workaround
We've implemented a subclass that filters empty text parts:
class FixedGoogleModel(GoogleModel):
async def _map_messages(self, messages):
system_instruction, contents = await super()._map_messages(messages)
# Filter out empty text parts and ensure valid structure
cleaned_contents = []
for content_item in contents:
if "parts" not in content_item:
cleaned_contents.append(content_item)
continue
original_parts = content_item.get("parts", [])
cleaned_parts = [
part for part in original_parts
if not (isinstance(part, dict) and part.get("text") == "" and len(part) == 1)
]
if not cleaned_parts:
cleaned_parts = [{"text": " "}]
content_item["parts"] = cleaned_parts
cleaned_contents.append(content_item)
if not cleaned_contents:
cleaned_contents = [{"role": "user", "parts": [{"text": " "}]}]
return system_instruction, cleaned_contentsThis workaround successfully resolves the 400 errors while maintaining API compatibility.
Example Code
import asyncio
from pydantic_ai import Agent
from pydantic_ai.models.google import GoogleModel
from pydantic_ai.providers.google import GoogleProvider
async def reproduce_bug():
# Set up GoogleModel with Vertex AI
provider = GoogleProvider(
vertexai=True,
location="us-central1" # or your preferred location
)
model = GoogleModel("gemini-pro-2.5", provider=provider)
# Create an agent
agent = Agent(model=model)
# This scenario can trigger empty message parts
# Case 1: Empty string message
try:
result = await agent.run("")
print("Empty string succeeded:", result)
except Exception as e:
print("Empty string failed:", str(e))
# Case 2: Message that becomes empty after processing
try:
result = await agent.run(" ") # Only whitespace
print("Whitespace succeeded:", result)
except Exception as e:
print("Whitespace failed:", str(e))
# Run the reproduction
if __name__ == "__main__":
asyncio.run(reproduce_bug())Python, Pydantic AI & LLM client version
Pydantic AI version: v0.3.1
Python version: 3.12
Provider: Google Vertex AI
Model: gemini-pro-2.5 (via GoogleModel/Vertex)